It wasn't the first time women had noticed this, but McClintock was intrigued. And it only made her more so when male researchers with whom she studied one summer at Jackson Laboratory in Bar Harbor, Maine, pooh-poohed the whole thing.

So when she got back to school in the fall, she recruited 135 women and kept track of their cycles and friendships. Within four months, women who were friends began to cycle together. The degree of synchronization increased over time, as if they were communicating via an undetectable chemical signal.

They were. Last week, after decades of work in rodents, McClintock proved it in a study in Nature that has scientists buzzing.

McClintock, a psychology professor at the University of Chicago, showed -- in research others describe as ``elegant'' and ``carefully controlled'' -- that odorless underarm secretions taken from some women then wiped under the noses of others can shorten or lengthen the recipients' cycles, depending on where in their own cycles the donors were when the secretions were taken.

This is the first solid evidence that pheromones -- chemicals emitted by one animal that exert a behavioral or physiological response in another -- can influence physiological responses among humans, though human pheromones per se have not been isolated.

Pheromones are fascinating.

By synchronizing ovulation, for instance, they may increase genetic diversity. Since males often mate according to size or dominance, if only one female is in heat at a time, chances are the dominant male will inseminate her. If many are in heat, more males have a chance to pass on their genes.

And if females give birth together, they nurse and care for each other's young, which has been shown to increase survival of the infants, says Julie Mennella , a former student of McClintock's and now a biopsychologist at the Monell Chemical Senses Center, a nonprofit research group in Philadelphia.

Pheromones also help animals mark territory and tell friend from foe. In some species, bedding from the cages of strange males can cause miscarriages if placed in the cages of pregnant females, says Charles Wysocki, a neuroscientist at Monell.

Other studies suggest that human mothers can identify their own newborns by the smell in T-shirts worn by the child and that infants prefer breast or armpit pads worn by their own mothers, says Israeli psychologist Aron Weller in a commentary accompanying McClintock's paper in Nature.

And pheromones clearly play a role in mating behavior.

When sexual attractant pheromones are put in traps, beetles can't help themselves; they flock to the stuff (and get stuck in the trap). When female pigs in heat get a whiff of a pheromone called androstenone from the saliva of males, they assume the mating position, whether a male is around or not.

When female hamsters are anesthetized and rubbed on the rear end with vaginal secretions, males mate with them even though they're out cold.

Nobody knows yet how pheromones affect human mating, if they do, though this has hardly cooled the ardor of hypesters -- check the Net -- touting products supposedly laced with pheromones. (One product is described as a ``perfectly legal sexual stimulant cleverly masked in a men's cologne that when unknowingly inhaled by any adult woman unblocks all restraints and fires up the raw animal sex drive in every woman.'')

Even if all that were true, the most interesting word in that purple prose is ``unknowingly,'' because pheromones can indeed act outside of conscious awareness, perhaps because of the specific nerve pathways through which pheromone signals travel.

In many species, the olfactory system consists of two parts. The main system, for the conscious detection of smells, involves a patch of nerve tissue high up in the nose. This tissue contains receptors that catch airborne molecules.

The neurons connect to the olfactory bulb, a lightbulb-shaped structure in the front of the brain that processes chemical signals. The bulb then passes the signals to higher centers in the cortex, or thinking part of the brain, where information is further processed and the animal becomes conscious of the scent, or at least acts as if it does.

But in many animals, there's also an accessory olfactory system, called the VNO, or vomeronasal organ, located lower in the nose closer to the mouth.

The intriguing thing about this organ is that its nerve fibers run to a different part of the olfactory bulb and from there go not to higher centers in the brain but to the amygala and hypothalamus, which can process signals without the animal's awareness. From the hypothalamus, signals then go to the pituitary gland, which controls ovulation and reproduction.

Because these signals don't wind up in the cortex, they may remain outside conscious awareness, says John Vandenbergh, a zoologist at North Carolina State University.

The big question is whether humans have something similar. Until recently, scientists thought not, or that if there is one, it is vestigial and disappears during fetal development.

But now, says Linda Buck, a neurobiologist at Harvard Medical School, the consensus is that there seems to be a structure that might be the human counterpart of the VNO, ``but no one has been able to demonstrate that it's connected to the brain.''

Buck and others have found genes in human DNA for two families of receptors that would respond to pheromones in a VNO system, but the genes are mutated and can't make functional receptors, she says.

With or without functioning VNO receptors, it's clear from McClintock's study of 29 women that humans can react to phermones somehow, ``either by using an unidentified part of the main olfactory system, or perhaps with a sixth sense with its own unique pathway,'' as McClintock puts it.

Armpit secretions obtained when the donors were in the first half of the menstrual cycle shortened recipients' cycles, by as much as 14 days. When secretions were taken when donors were ovulating, the recipients' ovulation was delayed and their cycles were lengthened, by as much as 12 days.

"It was like eau d'ovulation,'' says McClintock, and what makes the study ``quite elegant," adds Wysocki, is that "not only did the McClintock group show they could alter the cycle, they showed they could advance and retard it in the same women."

Weller, in the Nature commentary, calls the finding "ground-breaking," not least because pheromone-based drugs might be developed to help regulate menstrual cycles in infertile women. And "we may yet discover that other aspects of our behavior and physiology are affected by covert olfactory messages from other people during social interactions," he says.

Napoleon I certainly seemed to believe something like that, according to a famous message he is said to have sent to Josephine, the empress of France. "I return in three days," he wrote. "Don't bathe."

Granted, the effect of pheromones may be muted in modern society, where daily showers and deodorants are a mark of civilization. But the McClintock study shows we are more like other animals in our ability to process hidden chemical signals than some might like to think.

"It's like the Titanic in the sense that we can only see the tip of this iceberg,'' says Wysocki. ``We have only begun to realize how susceptible we may be to chemical communication among ourselves."